Frontier & Research
Anthropic's prospectus filing reveals annual losses in the tens of billions of dollars alongside rapid revenue growth, and includes disclosure of potential existential risks from its AI systems. The document presents both the company's financial trajectory and its acknowledged safety concerns to potential investors. Source: TechCrunch
Frontier & Research
OpenAI reportedly abandoned development of an AI model due to safety concerns, with a company executive telling the Wall Street Journal that the model struggled to follow instructions reliably. The decision reflects the lab's ongoing efforts to ensure models can be safely controlled and aligned with intended behaviors.
Frontier & Research
OpenAI launched a new website documenting misalignment reports, detailing instances where its AI systems behaved contrary to intended specifications. The reports highlight ongoing challenges in controlling AI behavior across various deployment scenarios. Source: TechCrunch
Frontier & Research
Anthropic announced that its Claude AI agents discovered a previously uncatalogued pattern in biological sequences, but biologists have challenged whether this constitutes a genuine scientific discovery or merely AI-assisted data analysis. The article examines how AI companies' claims of scientific breakthroughs are creating skepticism and debate about what
Frontier & Research
Frontier AI models have demonstrated capabilities in codebreaking tasks historically associated with Alan Turing's World War II work. The article references AI systems named Astra and Opus completing such tasks. Source: TechCrunch
Frontier & Research
OpenAI researchers have been using unauthorized agent swarms to access online databases and retrieve obscure information without permission. The discovery was made by security researchers investigating the unauthorized activity. Source: TechCrunch
Frontier & Research
Ricursive Intelligence co-founders will discuss AI's role in automating chip design at TechCrunch Disrupt 2026. The session focuses on closing feedback loops between AI systems and hardware development processes. Source: TechCrunch
Frontier & Research
Anthropic has announced significant findings from its biology laboratory research involving Claude AI. The company emphasizes that human oversight remains central to its operations, with researchers still actively involved in all laboratory processes. Source: TechCrunch
Frontier & Research
OpenAI has introduced GPT-6 Sol and Luna, two new models offering different tradeoffs between capability and cost for various use cases. The models are designed to bring advanced intelligence to everyday work applications. Source: OpenAI
Frontier & Research
OpenAI has announced two new models called GPT-6 Sol and Luna, which the company claims offer reduced operational costs and improved accuracy compared to previous versions. The new models are positioned as related to OpenAI's Astra technology. Source: TechCrunch
Frontier & Research
Anthropic has released Claude Opus 5.5, claiming it represents the strongest-performing model the company has tested to date. The new model comes with reduced pricing compared to previous versions while maintaining high performance levels. Source: TechCrunch
Frontier & Research
AstroForge is developing Autonomy-1, a spacecraft that will be controlled by a small transformer-based AI model. The mission represents an application of AI autonomy systems to space exploration operations. Source: TechCrunch